Cost to serve by channel: what a contact really costs

Workforce economics · ~7 minute read

“Chat is cheaper” — is it, though?

Channel-shift business cases lean on a comfortable assumption: that moving contacts from voice to chat, or from chat to self-service, saves money more or less automatically. Sometimes it does. But “cheaper” depends entirely on how you count, and the headline cost per contact that vendors and strategy decks quote usually counts the wrong things. A planner who can build an honest cost to serve — loaded, channel by channel — is one of the few people who can tell the business which channel-shift actually pays and which just moves the cost somewhere the spreadsheet didn’t look.

Building an honest number

The core of cost to serve is agent time: the handle time per contact multiplied by the fully loaded hourly rate — salary plus on-costs, plus the shrinkage and support layers that make an hour on the phones cost more than an hour’s pay. For chat you divide by concurrency, which is where the saving is supposed to come from — but concurrency also raises handle time, so the two partly cancel, and the real per-contact cost sits well above “voice divided by the number of chats.” Email and other deferred work carry their own handling and rework costs. Self-service looks almost free per contact, until you remember it doesn’t remove the hard contacts, it concentrates them — so the agent contacts that remain are dearer, and the apparent saving is partly a transfer. Add technology, supervision and overhead on top, and the true ranking of channels is often less flattering, and more interesting, than the headline.

Loaded cost per contact, by channel Voice Chat Email Self-serve * * near-free per contact — but it hardens the residual that still reaches agents Net the concurrency saving against the handle-time it adds before you rank them.
Cost to serve isn’t the list price of a channel; it’s the loaded, netted figure — and self-service’s cheapness comes with a bill that lands elsewhere.

Using cost to serve well

Once you have the loaded numbers, use them as a decision input, not a verdict. Channel-shift and deflection decisions should be made on true cost to serve, including the way self-service concentrates difficulty onto agents — otherwise you book a saving that reappears as a longer, harder queue. Hold cost to serve against satisfaction and resolution, because the cheapest channel that fails the customer generates a repeat contact and costs you twice; the goal is lowest cost per resolved need, not lowest cost per interaction. And keep the numbers current, because concurrency, automation and mix all move them. A planner who owns an honest cost to serve gives the business the one number most channel strategies are quietly missing — what a contact really costs, once everything is counted.

Pair this with self-service and containment, the cost-per-contact calculator, and the P&L view of planning.